Experimental investigation of the performance of industrial evaporator coils operating under frosting conditions

available at www.sciencedirect.com www.iifiir.org journal homepage: www.elsevier.com/locate/ijrefrig Experimental investigation of the performance of industrial evaporator coils operating under frosting conditions N.F. Aljuwayhel, D.T. Reindl*, S.A. Klein, G.F. Nellis University of Wisconsin-Madison, 1500 Engineering Drive, Madison, WI 53706, USA article info Article history: Received 10 March 2006 Received in revised form 19 May 2007 Accepted 21 May 2007 Published online 10 July 2007 Keywords: Refrigeration Industrial application Evaporator Finned tube Experiment Temperature Relative humidity Air Frosting abstract This paper describes a field experimental investigation of the effects of frost formation on the performance of a low-temperature large-scale evaporator coil used in industrial refrigeration systems. A series of experiments were conducted to determine the in situ coil cooling capacity of the evaporator over time as frost builds on its surfaces. Field-measured quantities include inlet and outlet air temperatures, inlet and outlet air relative humidity, and air volume flow rate. These measurements provide a baseline set of experimental data that can be used to validate numerical models of industrial evaporators operating under frosting conditions. ª 2007 Elsevier Ltd and IIR. All rights reserved. Etude expérimentale sur la performance des évaporateurs à serpentins industriels sous des conditions de givrage Mots clés :Réfrigération ; Application industrielle ; Évaporateur ; Tube aileté ; Expérimentation ; Température ; Humidité relative ; Air ; Gi 1. Introduction Refrigeration is an enabling technology in a wide range of applications from air conditioning for occupant comfort to freezing as required in food preservation. Evaporators are the critical component responsible for extracting heat from conditioned spaces or processes. The focus of this paper is on evaporators that cool air to temperatures below the freezing point of water. When an air-cooling evaporator operates at a temperature below the freezing point with a coincident entering * Corresponding author. Tel.: þ1 608 262 6381. E-mail address: dreindl@wisc.edu (D.T. Reindl). 0140-7007/$ see front matter ª 2007 Elsevier Ltd and IIR. All rights reserved. doi:10.1016/j.ijrefrig.2007.05.010

99 Nomenclature A f evaporator coil face area (m 2 ) i enthalpy (kj kg 1 ) _m mass flow, accumulation rate (kg s 1 ) m mass (kg) _q heat transfer rate (W) R generic measured result RH relative humidity ( ) T temperature (K) V velocity (m s 1 ) VR generic variable for uncertainty analysis d uncertainty r density (kg m 3 ) u humidity ratio Subscripts a air ave spatially averaged fr frost i inlet to the evaporator coil L location m mass, instrument o outlet from the evaporator coil t total air dew point temperature that is above the evaporator coil surface temperature, frost will form on the evaporator surface. The significance of the frost formation is twofold. First, the presence of frost reduces the ability of an evaporator fan to move air across the coil; as a result, the refrigeration capacity of the evaporator decreases as frost accumulates during operation. Second, the presence of the low conductivity frost layer represents an additional thermal resistance between the air and the refrigerant; thereby, decreasing the evaporator performance. For these reasons, accumulated frost must be periodically removed by the use of a defrost process. A large number of experimental and theoretical investigations have been reported relative to frost properties, the mechanisms of frost growth, and the heat transfer through frosted surfaces with simple geometries. For complex geometries, such as finned-tube heat exchangers, the available literature is more limited due in part to the large number of variables that affect frost growth on these surfaces. Most of the experimental studies of frosting finned-tube heat exchangers have been performed using relatively small scale heat exchangers in a laboratory environment. The heat exchangers studied in these previous experimental investigations were designed for applications other than industrial refrigeration (Stoecker, 1957; Hosoda and Uzuhashi, 1967; Sanders, 1974; Gatchilov and Ivanova, 1979; Senshu and Yasuda, 1990; Rite and Crawford, 1991; Lee et al., 1996; Yan et al., 2003). This paper describes a systematic field study performed on a large-scale evaporator coil used in industrial refrigeration operating at both low air and refrigerant temperatures. The in situ coil cooling capacity of an evaporator is measured over time as frost builds on its surfaces. These measurements provide a baseline set of experimental data that can be used to validate numerical models of industrial refrigeration systems. Table 1 Geometry and operating conditions of the coil used in the experiment Parameter Value Fin pitch (cm) 0.85 Face area (m 2 ) 8.23 Tube diameter (mm) 19.05 Tube length (m) 5.5 Number of fans 5 Fan power 2.33 at 30 F( 34 C) air temperature (kw) Rated air flow rate (m 3 min 1 ) 1699 Number of tubes 260 Number of tube row 10 Tube transverse pitch (mm) 57 Tube longitudinal pitch (mm) 44 Evaporation temperature ( C) 34.4 Coil temperature difference ( C) 5.6 Fin and tube material Aluminum Refrigerant Ammonia Evaporator coil type CPR-fed liquid overfeed 2. Experiment facility The coil selected for this experimental investigation is a liquid overfed evaporator that is installed in a penthouse on the roof of a storage freezer. The coil is used to maintain a space temperature of 29 C for the long-term storage of ice cream products. The geometric details of the coil used in the experiment and the nominal operating conditions are summarized in Table 1. Fig. 1 shows an isometric view of the penthouse with its overall dimensions. Fig. 2 shows the key components in the penthouse while Fig. 3 shows an isometric view of the penthouse components along with the air flow paths. During cooling mode operation, warehouse air enters the penthouse through a grate that is located in the penthouse floor perpendicular to the upstream coil face. The air is then drawn across the evaporator coil by five fans that are located in the penthouse floor perpendicular to the coil on the downstream side of the evaporator, as shown in Fig. 3. The fans discharge the air into a plenum through five round extension ducts, each with an opening of diameter 0.97 m. The plenum 2.73 m 2.8 m Access door 6.5 m Data acquisition cabinet Fig. 1 Schematic showing the outside dimensions of the penthouse enclosure.

100 0.97 m Penthouse 0.8 m 0.67 m 0.97 m Up stream T i RH i Evaporator Coil Down stream T o RH o V o 6.17 m 6.5 m 0.164 m Fig. 2 Plan view showing the internal dimensions of the penthouse enclosure. Air in Freezer Air out Fig. 4 Schematic showing locations for measurements taken during the experiments. then distributes the cold air to the freezer environment through 10 round exit ducts. The plenum is attached to the freezer ceiling. 3. Design of experiment Since the frosting process includes both sensible and latent air-side heat transfer, the evaporator coil heat transfer rate can be determined by applying an energy balance on the airside across the coil: _q t ¼ _m a ii;ave i o;ave (1) where _q t is the total evaporator coil heat transfer rate, i i;ave and i o;ave are the bulk enthalpies of the air in and out of the evaporator coil, respectively. Because the energy associated with the moisture that leaves the air stream and is deposited on the coil surface is small compared with the energy change of the moist air across the coil, it is neglected from the air-side energy balance. The air mass flow rate, _m a, can be computed according to: _m a ¼ r o V o;ave A f (2) where r o is the density of air downstream of the evaporator coil, V o;ave is the average velocity of air through the coil face, and A f is the evaporator coil face area. The rate of frost accumulation, _m fr, can be computed using the following equations: _m fr ¼ _m a ui;ave u o;ave (3) where u i;ave and u o;ave represent the average upstream and downstream humidity ratios, respectively. Fig. 4 shows the location of the air property measurements required to carry out the calculations represented by Eqs. (1) (3). Note that Fig. 4 does not identify the exact location or number of instruments used in the experiments; rather, it illustrates the state variables that result from averages of individual sensors. 4. Instrumentation and data acquisition A summary of the instrumentation used for measuring temperature, humidity, and flow is provided in the following sections. 4.1. Flow and temperature measurements Air inlet to the coil through the grid Coil The average air velocity leaving the evaporator (V o,ave ) is determined by taking the spatial average of the local air velocity 0.9m Down-stream of penthouse evaporator Uni-strut grid V 1 & T o1 fans 1.33m 1.09m 0.98m V 2 & T o2 0.66m V 3 & T o3 0.46m V 5 & T o5 V 4 & T o4 0.94m Cold air exit to the warehouse through plenum Fig. 3 Schematic of the experiment coil and the flow path of the air. 5.5m Fig. 5 Elevation view downstream of the evaporator coil showing the location of each air velocity transducer and thermistor.

101 0.3 m Thermistor Table 3 Summary of the estimated uncertainties of the measured variables Variable dvr m dvr L dvr s dvr overall Air velocity mass flow meter T i ( C) 0.1 0.2 0.01 0.22 T o ( C) 0.1 0.3 0.04 0.3 V (m s 1 ) 0.15 0.3 0.04 0.3 RH i (%) 2 0.2 0.017 2 RH o (%) 2 0.5 0.016 2 measured at five separate locations immediately downstream of the evaporator coil face using air velocity transducers. The five air velocity transducers (labeled V 1, V 2,., V 5 ) were placed 0.90 m apart laterally in order to divide the coil face into six equal areas and were offset 0.30 m from the downstream face of the coil (see Figs. 5 and 6). The bulk inlet and outlet air dry bulb temperatures are determined using nine thermistors. Five thermistors (labeled T o1, T o2,., T o5 ) are used to provide a measure of the average outlet air dry bulb temperature from the coil. The five thermistors are spaced diagonally across downstream coil face as shown in Fig. 5; each thermistor is mounted adjacent to a corresponding velocity flow meter as shown in Fig. 6. The remaining four thermistors (labeled T i1, T i2,., T i4 ) are used to obtain a bulk inlet air dry bulb temperature. Table 2 Average coil face parameters Parameter Equation Average air velocity Average air inlet relative humidity Average air outlet relative humidity V o;ave ¼ Average air inlet temperature T i;ave ¼ X4 X5 n¼1 V n!,5 (4) RH i;ave ¼ðRH i1 þ RH i2 Þ=2 (5) RH o;ave ¼ðRH o1 þ RH o2 Þ=2 (6) n¼1 Average air outlet temperature T o;ave ¼ X5 Heat shrink tube at connection with hook-up wire Fig. 6 Velocity sensor and the thermistor mounting detail. n¼1 T i;n!,4 (7) T o;n V n!, X 5 n¼1 V n (8) 4.2. Humidity measurements The average inlet and outlet air relative humidity were measured using four humidity sensors. Two humidity sensors each were mounted upstream and downstream of the evaporator. The humidity sensors used to measure the inlet air relative humidity (RH i1 and RH i2 ) are mounted above the defrost coil drain pan and 0.30 m from the upstream coil face. The two relative humidity sensors used to measure the outlet air relative humidity (RH o1 and RH o2 ) are attached at an elevation of 0.46 and 0.90 m and offsetby0.30mfromthedownstreamfaceofthecoil. 5. Experiment uncertainty, results and discussion The air velocity, dry bulb temperatures and relative humidity are measured using the experimental setup discussed in the previous section. Data were collected during cooling mode operation following the completion of a defrost cycle (i.e., starting with a clean coil) and each frosting condition test was repeated five times (individual tests are referred to as runs #1, 2, 3, 4 and 5). Measured data from each transducer are collected at 1 min intervals for a period of 42 h, except during runs #2 and #5 where the data collection spanned 31 and 22 h, respectively. The 1 min observations were averaged in order to develop 1 h time-average information for performance analysis of the cooling mode operation. A spatially averaged coil face air velocity (V ave ), inlet and outlet air relative humidity (RH i,ave and RH o,ave, respectively) and the average inlet temperature (T i,ave ) were obtained by calculating the mean value of the spatial 1-h interval averaged data. The bulk air dry bulb temperature on the downstream side of the evaporator (T o,ave ) was calculated using a velocityweighted average of the outlet temperatures, as shown in Table 2. The inlet and outlet air humidity ratios and enthalpies Table 4 Summary of the estimated uncertainties of the calculated results Calculated result, R dr i i;ave ðkj kg 1 Þ 0.12 i o;ave ðkj kg 1 Þ 0.15 u i;ave 5.0E 06 u o;ave 3.60E 06 _m fr ðkg h 1 Þ 0.83 m fr (%) 1.5 _q t ðkwþ 8.70

102 Fig. 7 Time-dependent spatially averaged air velocity on the downstream side of the evaporator (with time 0 being a frostfree coil). Fig. 8 Time-dependent measured air velocity of each velocity sensor and the calculated average air velocity downstream side of the evaporator; (a) run #1 and (b) run #3.

103 Fig. 9 Time-dependent spatially averaged inlet and outlet air temperatures throughout the experiment. were calculated using moist air property correlations based on the average temperature and relative humidity at these locations. 5.1. Uncertainty analysis The uncertainty associated with any of the variables measured in the experiment (referred to generically here as VR) can be related to three different factors (Table 3): 1. the design uncertainty of the instrument used in the experiment, dvr m, 2. the location of the instrument (i.e., spatial variations across the coil), dvr L, and 3. the random fluctuation and the scatter in the instrument readings, dvr s The overall uncertainty for each variable is calculated using the root sum square (RSS) technique which assumes that these sources of uncertainty are normally distributed and uncorrelated; the results are summarized in Table 4. ðdvr overall Þ 2 ¼ðdVR s Þ 2 þðdvr L Þ 2 þðdvr m Þ 2 (9) The 95% confidence level uncertainties of the calculated quantities are estimated using typical propagation of error techniques represented by Eq. (5) and the results are summarized in Table 4: vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ux 2 vr dr ¼ t dvr 2 overall;i (10) vvr i i where dr is the uncertainty of the measured result, R. 5.2. Experimental results The time variation of the average air velocity on the downstream side of the evaporator throughout each of the four experimental runs is shown in Fig. 7. Notice the general trend of Fig. 10 Time-dependent dry bulb temperature difference between the inlet and the outlet air flows.

104 Fig. 11 Time-dependent relative humidity of the inlet air. decreasing average face velocity with time which is attributable to the increase of air flow resistance associated with frost accumulation on the coil. Because a fixed speed fan is used to move air through the coil, the increase in air flow resistance results in a decrease in the flow rate. Fig. 8a and b illustrates the temporal variation of air velocity at the five spatial locations along the coil face for runs #1 and #3, respectively. Also shown is the average downstream air velocity calculated using Eq. (4). Fig. 8 shows that the velocity readings obtained from each of the separate velocity sensors are fairly consistent from run to run. For example, notice that the velocity reading associated with sensor V 3 (located in the center of the coil) is consistently higher than the velocity reading from the other four sensors at the beginning of each run. This difference was expected since the free air path across the evaporator at the elevation where the velocity sensor was mounted is slightly larger than the free air path at other locations across the evaporator. Fig. 8 also shows that the air flow rate is highly non-uniform downstream of the evaporator coil; this non-uniformity is mainly due to the design of the evaporator coil penthouse. The time variation of the bulk air inlet and outlet dry bulb temperature for each of the five different runs is shown in Fig. 9. The change in temperature experienced by the air as it passes through the coil is shown in Fig. 10 for each of the five runs. Figs. 9 and 10 show that the temperature drop of the air as it passes through the coil continuously increases as frost builds up; this is due to the continuous decrease in the average air velocity that was shown in Fig. 7. As the average air velocity decreases, the coil effectiveness increases, which causes a reduction in the leaving temperature of the air. It should be noted that the improved effectiveness is not sufficient to make up for the overall reduction in the coil refrigeration capacity associated with the reduction of the flow rate. It will be shown subsequently that the capacity of the evaporator continuously decreases over time. Figs. 11 and 12 illustrate the time variation of the average inlet air relative humidity and the average inlet and outlet Fig. 12 Time-dependent humidity ratio of the inlet and outlet air.

105 Fig. 13 Time-dependent frost mass accumulation rate. Fig. 14 Integrated mass accumulation of frost. Fig. 15 Time-dependent evaporator cooling capacity.

106 Table 5 Initial and final cooling rates for five experiment runs along with the percentage loss of the cooling rate Run Initial capacity (kw) Operating interval (h) Final capacity (kw) Capacity loss (kw) 1 121.5 8.7 42 94.3 8.7 27.2 2 120.4 8.7 31 89.5 8.7 30.9 3 114.5 8.7 42 93.9 8.7 20.6 4 110.4 8.7 42 90.1 8.7 20.3 5 120.8 8.7 22 110.7 8.7 10.1 air humidity ratios, respectively, during the five runs. Fig. 11 shows that the relative humidity is nearly constant during the duration of the experiment for the five runs. Fig. 12 shows that the humidity ratio of the exit air is always lower than that of the inlet air, as expected since moisture removed from the air builds up at the coil surface in the form of frost. Fig. 13 illustrates the rate of frost accumulation ð _m fr Þ as a function of time throughout the five runs. The rate of frost accumulation was calculated from a mass balance on the water carried by the air, Eq. (3). Fig. 14 shows the mass of the accumulated frost as a function of time ðm fr Þ for the five runs. Fig. 14 indicates that the mass of the accumulated frost grows, essentially linearly, with time; this is because the rate of frost accumulation changes only slightly during each run. A hot gas defrost cycle was carried out at the conclusion of run #5 and the quantity of melted frost was measured directly by collecting the condensate emanating from the coil drain. The total mass of condensate measured for run #5 was 188 kg. The estimate of accumulated frost that is obtained from integrating the frost accumulation rate over time is 200 3 kg as noted in Fig. 14. The discrepancy between these values is likely attributable to the experimental error associated with collecting and measuring the melt and the experimental error associated with calculating the total mass of the frost. The discrepancy could also be attributed to a portion of the frost being re-evaporated to the freezer space plus the portion of the melted frost that remains adhered to the coil surfaces and the drain pan at the end of the defrost cycle. Coley (1983) stated that during each defrost cycle, at least 15% of the ice sublimes back into the conditioned space to be removed again; however, the current experiment shows that for a coil being defrosted in a penthouse enclosure, the total accumulated frost that transfers back to the conditioned room as a latent load is less than 6%. If it is assumed that the thickness of the condensate film remaining on the coil surfaces after the defrost is about 1.0 mm, then the moisture mass remaining is approximately 9 kg which is equal to 4.5% of the total accumulated frost that remains inside the penthouse, leaving only 1.5% of the total accumulated frost that could possibly transfer back to the conditioned space as a latent load. Fig. 15 shows the evaporator cooling capacity as a function of time for the five runs calculated using Eq. (1). It can be seen from Fig. 15 that the evaporator coil cooling capacity decreases monotonically as expected, mainly due to the decreases in the air mass flow rate. The initial and the final calculated rates of cooling for each run along with the loss of cooling capacity are summarized in Table 5. 6. Conclusions An experiment has been conducted in order to measure the in situ coil cooling capacity of a large-scale industrial evaporator coil as frost builds up on its surface. Based on the experimental results, the following conclusions can be made: the accumulation of frost on the surfaces of the evaporator coil causes an increase in the resistance to air flow and an associated drop in the air mass flow rate through the evaporator coil; as the mass flow rate of the air passing through the evaporator coil decreases, the temperature drop of the air as it passes through the coil will increase as a result of the leaving air temperature decreasing; and the evaporator cooling capacity decreases monotonically, mainly due to the decreases in the air mass flow rate. Acknowledgments This project was financially supported by Wells Dairy Inc., the Kuwait University Mechanical Engineering Department, and University of Wisconsin Industrial Refrigeration Consortium at Madison, WI. references Coley, M.B., 1983. The cost of frost. ASHRAE Journal 82 (9). Gatchilov, T.S., Ivanova, V.S., 1979. Characteristics of the frost formed on the surface of the finned air coolers. In: XV International Congress of Refrigeration, Venice, Paper # B2-71. Hosoda, T., Uzuhashi, H., 1967. Effects of frost on the heat transfer coefficient. Hitachi Review 16 (6). Lee, T.H., Lee, K.S., Kim, W.S., 1996. The effects of frost formation in a flat plate-finned tube heat exchanger. In: International Refrigeration Conference at Purdue. Rite, R.W., Crawford, R.R., 1991. The effect of frost accumulation on the performance of domestic refrigerator-freezer finned tube evaporator coil. ASHRAE Transactions 97 (Part 2). Stoecker, W.F., 1957. How frost formation on coils effects refrigeration systems. Refrigerating Engineering 65 (2). Sanders, C.T., 1974. The influence of frost formation and defrosting on the performance of air coolers. Ph.D. dissertation of Delf University, Netherlands. Senshu, T., Yasuda, H., 1990. Heat pump performance under frosting condition: part I heat and mass transfer on crossfinned tube heat exchangers under frosting conditions. ASHRAE Transactions 96 (Part 1). Yan, W.M., Li, H.Y., Wu, Y.J., Lin, J.Y., Chang, W.R., 2003. Performance of finned tube heat exchangers operating under frosting conditions. International Journal of Heat and Mass Transfer 46 (6).